Table 3.

Results of ML analysis (eight methods) of BPW indices for discriminating between robust and possible sarcopenia (dynapenia + presarcopenia + sarcopenia).

Accuracy	SVM	MLP	GNB	DT	RF	LR	LDA	KNN
1	53.40	61.06	49.04	53.92	59.08	58.44	71.56	55.95
2	62.21	64.56	53.46	56.42	64.56	60.73	67.65	59.05
3	57.21	55.55	51.23	56.97	61.71	55.24	71.64	60.11
Average	57.60	60.39	51.25	55.77	61.79	58.14	*70.28	58.37

Sensitivity	SVM	MLP	GNB	DT	RF	LR	LDA	KNN
1	0.34	0.45	0.39	0.4	0.45	0.42	0.66	0.39
2	0.4	0.53	0.34	0.56	0.44	0.43	0.55	0.32
3	0.47	0.61	0.42	0.62	0.65	0.45	0.64	0.48
Average	0.4	0.53	0.38	0.52	0.51	0.43	*0.61	0.39

Specificity	SVM	MLP	GNB	DT	RF	LR	LDA	KNN
1	0.73	0.77	0.59	0.67	0.73	0.75	0.76	0.73
2	0.78	0.73	0.68	0.56	0.8	0.73	0.76	0.79
3	0.64	0.5	0.57	0.52	0.58	0.62	0.76	0.68
Average	0.716	0.66	0.613	0.58	0.7	0.7	*0.76	0.73

AUC	SVM	MLP	GNB	DT	RF	LR	LDA	KNN
1	0.54	0.61	0.49	0.54	0.59	0.59	0.72	0.56
2	0.60	0.63	0.51	0.56	0.62	0.59	0.66	0.56
3	0.56	0.56	0.50	0.58	0.62	0.54	0.71	0.59
Average	0.56	0.60	0.50	0.56	0.61	0.57	*0.70	0.57

The accuracy is in %. Values are for threefold cross-validation. Asterisks indicate the highest average value. Accuracy, sensitivity, specificity and AUC were all highest for LDA.